There was once optimism among philosophers that functionalism could give a complete account of the mind. Today philosophers are a lot less sure of this due mostly to the arguments expounded by Block in his now classic “Troubles with Functionalism,” (Block 1993), as well as his later “Inverted Earth” (Block 1997), where he argued that functionalism cannot account for qualitative states. There are at least two strategies that one could take in response to Block’s arguments. First there is what Block has called the ‘containment response’. One gives up on qualitative states but holds that beliefs, indeed thoughts in general, can still be given a purely functional account. This sometimes takes the form of ‘belief box’ talk. One says that p is in one’s belief box and this is supposed to be shorthand for ‘p is playing the belief function,’ where this means that p has characteristic connections to characteristic inputs and outputs.

This is the strategy that Fodor has adopted for years. I think it is well known that he endorses a functional account of what beliefs are (though this is not to say that they have functional definitions) and that this is part and parcel of the representational theory of mind. He has recently gone on to argue that in order for the representational theory of mind to be successful it needs to be able to provide an account of what concepts are. Where at the common sense level concepts are the components out of which beliefs are made. So, on his usage, the belief that grass is green is made up of the concepts GRASS, IS and GREEN. The reason that it is the belief that grass is green (as opposed to the belief that water is wet) is because of the concepts which are in the belief box (are playing the belief role).  It is also well known that he has argued that of all the theories that are out and about in cognitive science none of them stand up to the various requirements that things like compositionality and systematicity require. This has led him to formulate conceptual atomism. Sadly, though there is a problem. Conceptual atomism is not compatible with a functional account of what the attitude part of the propositional attitudes consists in. Since Fodor thinks that atomism is the only theory of concepts compatible with the representational theory of mind, this is a big problem indeed. First I will rehearse the inverted qualia argument and then argue that a version of this argument can be run on beliefs if atomism is true.

 The inverted qualia argument, you will remember, goes as follows. We imagine two twins, let’s call them Pat and Tap. Now Tap has special lenses installed in his eyes at birth. These are the infamous ‘inverting lenses’ which cause the person in whom they are implanted to have inverted qualitative experiences. Thus Tap sees what Pat sees when looking at fire trucks (i.e. red) while he (Tap) is looking at grass (i.e. green) and vice versa. These children then grow up as usual. By the time they are in High School the two twins function identically. They use all the color words correctly, each calling red things red and green things green but one of them sees what we call green when looking at red things. They have inverted qualitative states but identical functional states and this suggests that qualitative character is not captured by the functional description of the twins. Once one has gone this far it is a short step to the absent qualia argument, which just supposes that we might have the functional state without any qualitative aspect to it at all. If one does not want to take the containment response then one can try and show that absent qualia are impossible and that will help to save the theory. This is the strategy that Shoemaker famously takes. He argues that the qualitative states will have many connection to belief states tsuch that we would not have the releveant kinds of belief states in the absence of the qualitative state.

It is generally taken for granted that the propositional attitudes are immune to this kind of argument, partly due to the alleged fact that these states do not have any qualitative character associated with them.  Block sums up the common sense view in Troubles with Functionalism when he says

…it is very hard to see how to make sense of the analog of spectrum inversion with respect to nonqualitative states. Imagine a pair of persons one of whom believes that p is true and that q is false while the other believes that q is true and that p is false. Could these two persons be functionally equivalent? It is hard to see how they could. Indeed, it is hard to see how two persons could have only this difference in beliefs and yet there be no possible circumstance in which this belief difference would reveal itself in different behavior. (p. 247)

Suppose that P is ‘dogs are nice’ and Q is ‘cats are nice’ then Pat would have to believe that dogs are nice and that cats are not nice while Tap would believe that cats are nice and that dogs are not nice. It is hard to see how this difference in belief would not result in some difference in behavior regarding cats and dogs. If there are differences in their behavior then these two are not functionally identical.

But then in the footnote to this passage Block admits that there is a sense in which we can have inverted beliefs. He asks us to imagine two distinct afflictions. One is the lenses that we are familiar with from the inverted qualia argument; this he calls ‘Stimulus Switching.’ A person wearing these lens will calls red things ‘green’ because he (falsely) believes them to be green. The second ailment, called ‘Word Switching’ is an ailment where the victim simply uses the incorrect (but opposite) words for the colors. This person, then, calls red things ‘green’ but has normal color beliefs; in other words he will call something ‘green’ but only accidentally, he really means red, and he believes that the object is red.

Now suppose that a victim of Stimulus Switching suddenly becomes a victim of Word Switching…He speaks normally, applying ‘green’ to green patches and ‘red’ to red patches. Indeed he is functionally normal. But his beliefs are just as abnormal as they were before he became a victim of Word switching…So two people can be functionally the same, yet have incompatible beliefs. Hence the inverted qualia problem infects belief as well as qualia (though presumably only qualitative belief).

To illustrate this again imagine our two twins: When Pat and Tap are both looking at a red apple, both will say that it looks red and both will behave in just the same ways towards the apple as would the other. Except that Pat believes that the apple is red while Tap believes that the apple is green. Calling ‘the apple is red’ p and ‘the apple is green’ q we can see that Pat believes that p is true and q is false while Tap believes that p is false and q is true. So this really is a case of belief inversion in the way that Block says is hard to imagine happening. This seems to me to be the same kind of thing that happened to Locke when he imagined his missing shade of blue but then goes on to dismiss it as unimportant.

What does Block mean when he says ‘presumably only qualitative belief’? He (presumably) means those beliefs that are connected to qualitative states, and this would seem to block Shoemaker’s defense of functionalism. This will include more than just beliefs about colors. It will include all of our perceptual beliefs as well as any beliefs that stem from them. So we cannot define qualitative similarity in functional terms in the way that Shoemaker needs. Shoemaker’s response depends on it being the case that to believe that we are in pain and yet not actually be in pain cannot happen. But there is some reason to think that this may be possible. And the fact that we can have massive perceptual belief inversion means that the connection to other states cannot help us to pin down the pain state functionally.

As I mentioned earlier, Fodor argues that for the representational theory of mind to work it needs conceptual atomism, so let me briefly say what that is. He has argued that anyone who endorses a RTM has to endorse conceptual atomism. Concepts are primitive and acquire their content via some ‘locking relation’ to things in the world. There are two choices for two the ‘locking relation’. One is the Causal/historical kind that is taken by Kripke, Devitt, and Millikian. Fodor has argued that these kinds of accounts can’t provide sufficient condition for concept acquisition. As he puts it, ‘causally interacting with doorknobs’ could not be enough to acquire the concept something in the head must have happened, presumably in the head! Since he thinks that it can’t be learning there is only one option left. Concepts must work like appearance properties.Red things are the things that produce in us a certain predetermined qualitative state. Nothing fishy here, standard Empiricism, really; just as red ‘triggers’ a preset state in a sensory space, so too with doorknobs. Being a doorknob is being the kind of thing that creatures with minds like ours ‘resonate’ to. This is his controversial claim that all concepts are innate

Now we can see why atomism is subject to inversion argument. Let’s again take P to be ‘cats are nice’ and Q to be ‘dogs are nice’. If MOST concepts are appearance concepts the we can run Block’s argument on ‘cat’ and ‘dog’ instead of ‘red’ and ‘green’. We imagine a device that when worn inverts the perception of cats and dogs. Thus when Tap wears the device he will see a dog where Pat sees a cat and a cat where Pat sees a dog. Imagine Pat and Tap both looking at my wiener dog Frankie. This is exactly analogous to the case before. These two are functionally identical people yet one believes that Frankie is a dog and not a cat while the other believes that Frankie is a cat and not a dog. So functionalism cannot account for intentional states if concepts are appearance properties.

So the situation is that if one thinks that the representational theory of mind is important and that it would be nice if something like that could work then one is committed to atomism. But atomism means that functionalism about the attitudes can’t be right.